Transdifferentiation of Chicken Embryonic Cells Into Muscle Cells by the 3' Untranslated Region of Muscle Tropomyosin THOMAS J

Proc. Natl. Acad. Sci. USA Vol. 92, pp. 7520-7524, August 1995 Cell Biology

Transdifferentiation of chicken embryonic cells into muscle cells by the 3' untranslated region of muscle tropomyosin THOMAS J. L'ECUYER*, PAUL C. TOMPACHt, ERIC MORRISt, AND ALICE B. FULTON*§ Departments of tBiochemistry, tOral and Maxillofacial Surgery, and *Pediatrics, University of Iowa, Iowa City, IA 52242 Communicated by Sheldon Penman, Massachusetts Institute of Technology, Cambridge, MA, May 3, 1995

ABSTRACT Transfection with a plasmid encoding the 3' Drosophila pattern formation is influenced by the 3' UTRs untranslated region (3' UTR) of skeletal muscle tropomyosin of nanos (23) and bicoid (24). The posterior determinant induces chicken embryonic fibroblasts to express skeletal nanos, important in abdominal segmentation, exerts its effect tropomyosin. Such cells become spindle shaped, fuse, and by its 3' UTR inhibiting maternal hb gene expression. Dro- express titin, a marker of striated muscle differentiation. sophila embryos acquire anterior-position determination by a Skeletal muscle tropomyosin and titin organize in sarcomeric gradient of bicoid protein; this gradient requires the 3' UTR arrays. When the tropomyosin 3' UTR is expressed in osteo- of bicoid mRNA. blasts, less skeletal muscle tropomyosin is expressed, and titin A 3' UTR can affect mineral metabolism. Eukaryotic sel- expression is delayed. Some transfected osteoblasts become enoproteins have sites within the 3' UTR that are critical for spindle shaped but do not fuse nor organize these proteins into selenocysteine incorporation in the coding region (25). An sarcomeres. Transfected cells expressing muscle tropomyosin iron response element within the ferritin and transferrin organize muscle and nonmuscle isoforms into the same struc- receptor transcripts allows the expression of these genes to be tures. Thus, the skeletal muscle tropomyosin 3' UTR induces controlled by iron availability (26, 27). The iron response transdifferentiation into a striated muscle phenotype in a element for transferrin receptor, to which a protein binds, is cell-type-specific context. found within the 3' UTR of its transcript. A different function for 3' UTRs is revealed by localization The 3' untranslated region (3' UTR) of transcribed RNA has of actin transcripts in myoblasts, which is determined by the 3' recently been demonstrated to have important biological func- UTR of its mRNA; this region may function as a "zip code" tions. The 3' UTRs of two important protooncogenes, c-myc (28). The 3' UTR also determines localization in vivo for and c-fos, have AU-rich elements that act to destabilize the transcripts of the chloroplast genes rbcl and psaB in Chlamy- mRNA domonas (29). full-length by facilitating the removal of the poly(A) The most surprising consequence of a 3' UTR was observed site; in this fashion the 3' UTRs of protooncogenes control the for the muscle structural proteins troponin I, tropomyosin, and amount of their own mRNA that is available for translation cardiac actin. The 3' UTRs of these transcripts have been (1-6). Similarly, the cytokine granulocyte/macrophage colo- shown by genetic complementation to permit differentiation in ny-stimulating factor (GM-CSF) has a region within its 3' UTR a mutant myoblastic cell line that is defective in differentiation that causes degradation of the message (7, 8). Since the (30). Wild-type muscle cells were augmented in their differ- half-life of this mRNA controls its gene expression level, the entiation by introduction of these 3' UTR sequences, and 3' UTR has a profound influence on GM-CSF gene expres- proliferation of 10T'/2 fibroblasts was suppressed, suggesting sion. Interleukin 1 increases GM-CSF expression by stabilizing that the 3' UTRs of certain differentiation-specific RNAs are its mRNA, perhaps by binding to the 3' UTR of GM-CSF trans-acting regulators in a feedback loop that can inhibit mRNA. The 3' UTR of human papillomavirus acts to desta- proliferation and promote differentiation. The 3' UTR of bilize late mRNAs as well, negatively influencing its own gene a-tropomyosin also suppressed tumor growth in mutant myo- expression (9). Another example of 3' UTRs influencing gene genic cells that form tumors in mice, further suggesting that expression by destabilizing their mRNAs occurs in a highly untranslated RNAs can function as regulators of cell prolif- unstable set of transcripts in soybeans, the auxin-up RNAs eration (31). The work to be described below extends these last (10). In Xenopus oocytes, the 3' UTR of a transcript called observations by showing that 3' UTRs can transdifferentiate XRh box 2B causes endonuclease activity ofspecific nucleotide primary chicken fibroblasts into muscle cells. sequences, providing another example of how this region can negatively influence gene expression, but of sequences other than its own (11). Numerous other transcripts have their MATERIALS AND METHODS stability decreased through elements in the 3' UTR (12, 13). Cell Culture. Fibroblasts were cultured from day 12 chicken Expression is also controlled by decreases in translatability (14, embryos by autodigestion of skin as described (32). Cells were 15). The role of the 3' UTR on translation has recently been grown on 100-mm dishes in Dulbecco's modified Eagle's reviewed (16, 17). medium (DMEM), supplemented with 10% horse serum, In other cases, a 3' UTR has a positive influence on gene 2.5% chicken embryonic extract, nonessential amino acids, expression. The 3' UTR of tobacco mosaic virus mRNA forms minimum essential medium vitamins, Hepes as a buffer, and a tertiary structure that includes a pseudoknot. Adding this 3' gentamicin until they reached confluency, at which time the UTR to foreign transcripts dramatically increases their expres- culture was passed by trypsinization onto collagen-coated glass sion in prokaryotes and eukaryotes (18, 19), by increases in coverslips at a density of 1.25 x 105 cells per dish. After cells translational efficiency and mRNA stability. Other mRNAs reached 80-90% confluency on coverslips, fibroblasts were also show increases in stability (20) or translational efficiency used for transfection experiments. Osteoblast cultures were with given 3' UTRs (21, 22). prepared by obtaining calvarial explants from day 12 chicken embryos. The explants were placed in 35-mm tissue culture The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in Abbreviation: 3' UTR, 3' untranslated region. accordance with 18 U.S.C. §1734 solely to indicate this fact. §To whom reprint requests should be addressed. 7520 Downloaded by guest on September 27, 2021 Cell Biology: L'Ecuyer et aL Proc. Natl. Acad. Sci. USA 92 (1995) 7521 plates with culture medium consisting of high glucose DMEM Nordic Immunologicals (Tilburg, The Netherlands). An anti- supplemented with 10% fetal bovine serum and containing 50 body to f3-galactosidase was obtained from Promega and used units of penicillin per ml, 50 jig of streptomycin per ml, 2 ,ug at 1:500 dilution. of amphotericin per ml, 50 ,ug of ascorbic acid per ml, 250 ,ug of glutamine per ml, and 20 mM Hepes. Outgrowths of cells from the bone fragments appeared within 4-5 days and formed RESULTS confluent monolayers at 7-10 days. After the bone explants Tropomyosin isoforms assemble by different pathways in were removed, the confluent cells were trypsinized and plated skeletal muscle and in nonmuscle cells (40). When we sought on collagen-coated coverslips in 35-mm culture plates at 1 x to examine the assembly of muscle-specific tropomyosin in 105 cells per ml. All cells are maintained at 37°C in a chicken embryonic fibroblasts by transfecting the cells with a humidified incubator with 5% CO2. plasmid encoding chicken skeletal muscle a-tropomyosin, an Production of Construct for Transfection. A plasmid con- unexpected result was obtained. Passaged chicken embryonic taining the avian skeletal a-tropomyosin cDNA was provided fibroblasts that expressed skeletal tropomyosin from a plasmid by S. Hitchcock-DeGregori (33) and was cloned into the during a transient transfection became spindle shaped and eukaryotic expression vector pCMV5 (34). The 3' UTR of this elongated (Fig. 1A). Many of the spindle-shaped cells that cDNA was amplified by designing PCR primers to the 5' and express skeletal muscle tropomyosin also express titin, the 3 3' ends of this sequence [excluding the poly(A)denylation mD marker of striated muscle differentiation (Fig. 1B). Tro- sequence] and performing PCR using full-length skeletal pomyosin and titin were found in fibrillar longitudinal strands, a-tropomyosin as a template. Nucleotides encoding unique and in some cells the proteins were organized in sarcomeric restriction sites were added to either end of the primers to arrays, as early as 24 hr after transfection. In areas where allow directional cloning into a vector (pCMV5) cut with the several cells expressing tropomyosin were close, they appeared same restriction enzymes. A 300-bp fragment was released to have fused, as multinucleated cells with as many as five from DNA purified from bacteria transformed with this nuclei were observed. Cultures allowed to express muscle construct after digestion with the unique restriction enzymes tropomyosin for 72 hr showed a higher frequency of multinu- flanking the cloning site, as expected. All cloning procedures cleated cells with titin in were adapted from described protocols (35). Control trans- periodic arrays. Dishes containing fections were performed with the vector pCMV5 alone and cells that had been mock-transfected showed no cells express- with pRSV-,Bgal, a construct encoding the lacZ gene (gift of R. ing muscle tropomyosin or titin. By all criteria used, the Singer; ref. 28), allowing the determination of whether cells fibroblasts had transdifferentiated into striated muscle cells. exposed to transfection cocktail take up and express foreign Numerous examples point to connections between cell DNA. shape and gene expression (41), so it was possible that this DNA Transfection. Several transfection methods were com- transdifferentiation was a consequence of a cell shape change pared. Best results were obtained when cells were transfected induced by expression of skeletal tropomyosin. However, with small unilamellar liposomes containing LipofectAmine genetic screens for regulators of muscle development detected (GIBCO/BRL) using published procedures (36). Cells were several muscle-specific 3' UTRs that complemented the defi- cultured as above until they were 80-90% confluent. For each ciencies of a differentiation-minus muscle strain (30). To 35-mm dish, 5 ,ug of DNA was diluted to 100 ,ul in Opti-MEM distinguish between these two possibilities, we transfected (Life Technologies, Gaithersberg, MD) supplemented with 55 embryonic fibroblasts with a plasmid expressing only the 3' ,uM 2-mercaptoethanol (2-ME) (GIBCO/BRL), and 10 ,ul of UTR of skeletal muscle tropomyosin. Cells in cultures trans- LipofectAmine was diluted to 100 1.l in Opti-MEM/2-ME. fected with this construct also became spindle shaped (Fig. The DNA and reagent dilutions were then combined in a 2A4). Spindle-shaped cells expressed skeletal muscle tropomy- polystyrene tube and allowed to stand for 20 min at room osin; many also expressed titin, often in periodic arrays. temperature. Cells were washed twice with Opti-MEM/2-ME, Multinucleated cells expressing titin often had the overall and 800 ,ul of serum-free Opti-MEM/2-ME medium was morphology of young myotubes (Fig. 2B). Another marker of added to the cultures. The 200 ,ul of reagent-DNA complex muscle differentiation, muscle-specific myosin heavy chain, was added to the cultures, and the cells were incubated 6-8 hr was also detected in these cells (Fig. 2C). When cells expressing before adding an equal volume of complete growth medium titin or myosin were in proximity they apparently had fused, as with 20% serum, bringing the final serum concentration to multinucleated cells were observed (Fig. 2 B and C). All of the 10%. Cells received fresh growth medium at 24 hr and were criteria for transdifferentiation of fibroblasts observed after processed for immunofluorescent staining 24-72 hr after transfection. Immunofluorescent Staining. Tropomyosin protein is detected in cultured cells by using antibodies specific for muscle tropomyosin (37) and nonmuscle tropomyosin (T4780; Sigma) after transfection, similar to a published procedure (38). Titin and myosin heavy chain proteins are detected with antibodies specific for muscle titin (39) and myosin heavy chain (MF20, a kind gift of D. Fischman, Cornell University). Cells were rinsed with phosphate-buffered saline (PBS) at room temperature before being fixed for 20 min at -20°C in methanol with 3.7% formaldehyde. Cells were rinsed with PBS, primed with PBS and 0.1% bovine serum albumin (BSA), and incubated with a mixture -of T4780 (Sigma) and CH1 (obtained from FIG. 1. Passaged chicken embryonic fibroblasts stained to reveal University of Iowa Hybridoma Center) in concentrations of skeletal tropomyosin expressed from a plasmid encoding full-length 1:200 and in PBS and BSA at for avian striated a-tropomyosin during a transient transfection. (A) Cells 1:100, respectively, 0.1% 37°C become spindle shaped and are stained for tropomyosin, found in 30 min. After rinsing off the first antibody mixture with PBS longitudinal fibrillar strands. (B) Cells that express skeletal muscle and repriming, fluorescently tagged second antibodies were tropomyosin also express titin, the 3 mD marker of striated muscle added at dilutions of 1:100 and incubation was performed at differentiation; in some cells titin is organized in sarcomeric arrays, as 37°C for 30 min. Second antibodies used were rhodamine- early as 24 hr after transfection. Cell shown here is stained for titin tagged anti-IgG2a and fluorescein-tagged anti-IgGl from only. (x1265.) Downloaded by guest on September 27, 2021 7522 Cell Biology: L'Ecuyer et al. Proc. Natl. Acad. Sci. USA 92 (1995) also expressed titin (Fig. 3B), which, like the muscle tropomyosin, showed little organization. In addition, cells expressing titin and tropomyosin appeared vacuolated and were never multinucleated, although clusters of three to six cells were often seen. It appeared that although the 3' UTR of tropomyosin could evoke several components of muscle differentiation from these cells, the osteoblasts were limited in their response to the 3' UTR and failed to progress as far along the path of muscle differentiation. The usual sequence of development in muscle involves turning off nonmuscle genes as well as turning on muscle- specific ones. For tropomyosin, this change in expression involves changing the use of splice sites during mRNA pro- cessing. When cultures were stained with antibodies specific for muscle or for nonmuscle isoforms, cells that expressed the muscle isoform still contained nonmuscle tropomyosin (Fig. 4A). In Fig. 4, the lefthand column (A, D, and G) shows cells stained for nonmuscle tropomyosin; the center column (B, E, and H) shows the same cells stained for muscle tropomyosin; FIG. 2. Embryonic fibroblasts transfected with a plasmid express- the righthand column (C, F, and I) shows the images of muscle ing only the 3' UTR of skeletal muscle a-tropomyosin. (A) Cells and nonmuscle tropomyosin merged. In Fig. 4, the top row transfected with this construct also become spindle shaped; staining (A-C) shows a transfected fibroblast expressing nonmuscle for tropomyosin reveals skeletal muscle tropomyosin in longitudinal and muscle tropomyosin; the middle row (D-F) shows a and periodic arrays. (B) Multinucleated cells stained for titin contain it in periodic arrays. (C) Another marker of muscle differentiation, fibroblast in the same culture expressing only nonmuscle muscle-specific myosin heavy chain, is detected in these multinucle- tropomyosin; the bottom row (G-I) shows a transfected os- ated cells. (X890.) teoblast expressing muscle and nonmuscle tropomyosin and several nontransfected cells expressing only nonmuscle tropo- transfection with the full-length muscle tropomyosin were also myosin. In fibroblasts and osteoblasts, the two tropomyosin observed after transfection with the 3' UTR of muscle tropo- isoforms are found extensively colocalized. However, in fibro- myosin alone. blasts expressing the muscle isoform, tropomyosin is found in Fibroblasts and myoblasts derive from mesenchymal lin- longitudinal fibers and periodic arrays. In osteoblasts and in eages. The ability of the 3' UTR to transdifferentiate fibro- fibroblasts expressing only nonmuscle tropomyosin, the tro- blasts may reflect the closeness of fibroblasts to myoblasts in pomyosin isoforms are found in stress fibers that are distrib- developmental pathways. To determine how broadly the 3' uted through the cell. UTR of muscle tropomyosin could act to transdifferentiate Several controls were performed. These included transfec- cells, the tropomyosin 3' UTR was transfected into embryonic tion with the viral vector lacking the tropomyosin 3' UTR chicken osteoblasts. At 24 hr after transfection, many osteo- insert; in this case, no cells expressed muscle tropomyosin (Fig. blasts expressed muscle tropomyosin but remained flat and 5 A and B). An additional control to confirm that skeletal isodiametric (Fig. 34). The muscle tropomyosin concentrated tropomyosin expression was induced by transfection was to along stress fibers but showed little periodicity. Cells with this transfect cells with plasmids encoding tropomyosin 3' UTR shape did not express titin. At 48 hr after transfection, many and 3-galactosidase simultaneously. When such cells were osteoblasts that expressed muscle tropomyosin were spindle stained for tropomyosin and f3-galactosidase, most cells posi- shaped. The muscle tropomyosin in these cells showed little tive for tropomyosin were also positive for 03-galactosidase organization. Spindle-shaped cells and some isodiametric cells (Fig. 4 C and D), confirming that these cells were transfected. DISCUSSION AND CONCLUSIONS Regulation by the Tropomyosin 3' UTR. A genetic screen developed by Rastinejad and Blau (30) revealed that four muscle-specific proteins had 3' UTRs in their mRNA that could complement the deficiencies of a mutant muscle line unable to complete differentiation. The investigators also examined the effect of the tropomyosin 3' UTR in 10T½/2 fibroblasts; when these immortal fibroblastic cells expressed the tropomyosin 3' UTR, they stopped dividing but did not progress further along the path of muscle development. Ex- pression of the tropomyosin 3' UTR also reversed the phenotype of a transformed cell line, restoring anchorage depen- dence and suppressing tumor formation (42). Although several possible mechanisms for these effects were discussed, none has yet been identified. The investigators proposed calling such RNAs "riboregulators." The observation of complementation by Rastinejad and Blau (30) was surprising. Those investigators observed growth FIG. 3. Osteoblasts 24 hr after transfection with the tropomyosin inhibition when cells to the 3' UTR 3' UTR. (A) Cells stained for muscle tropomyosin reveal diffuse and only they exposed 10T/2 linearly organized, but not periodic, muscle tropomyosin; cells remain of mouse skeletal tropomyosin. That the fibroblasts examined flat and isodiametric. (B) Isodiametric cells stained for titin display the in the current study transdifferentiated into muscle may reflect muscle protein, which, like muscle tropomyosin, is not periodic. in part the euploid state of the embryonic chicken fibroblasts. (x680.) Other investigators have obtained better muscle differentia- Downloaded by guest on September 27, 2021 Cell Biology: L'Ecuyer et al. Proc. Natl. Acad. Sci. USA 92 (1995) 7523

FIG. 4. Organization of tropomyosin isoforms in transfected fibroblasts and osteoblasts. (A, D, and G) Cells stained for nonmuscle tropomyosin. (B, E, and H) Same cells stained for muscle tropomyosin. (C, F, and I) Images of muscle and nonmuscle tropomyosin merged. (A-C) A transfected fibroblast expressing muscle and nonmuscle tropomyosin. (D-F) A fibroblast in the same transfected culture expressing only nonmuscle tropomyosin. (G-I) A transfected osteoblast expressing muscle and nonmuscle tropomyosin. In fibroblasts and osteoblasts, the two tropomyosin isoforms are found extensively colocalized, as displayed in the merged images. However, in fibroblasts expressing the muscle isoform, tropomyosin is found in longitudinal fibers and periodic arrays. In osteoblasts, the tropomyosin isoforms are found colocalized in stress fibers that are distributed through the cell. Thus, the endogenous nonmuscle tropomyosin per- sists in transfected cells but is organized differently in fibroblasts than in osteoblasts. (x 1000.) tion with myoD by expressing it in primary cultures (43), which usually stable cells, ones that replicate principally in response are presumably euploid; 10T/2 cells are not. to wounding or damage, rather than postreplicative cells, such More than euploidy of the cells is required for transdiffer- as neurons, or ones continually renewing a tissue, such as entiation, however. The osteoblasts tested here began express- keratinocytes. Cues for transdifferentiation are usually com- ing titin and tropomyosin but failed to organize the muscle ponents of the extracellular matrix [collagen or laminin, for proteins or to fuse. The osteoblasts are also euploid, so this example (45)] or growth factors [including fibroblast growth difference in response presumably reflects differences in de- factor, transforming growth factor f, and insulin-like growth velopmental potential between fibroblasts and osteoblasts. factor I (48)]. Some cells transdifferentiate in response to Such differences are observed in other cases of transdifferen- soluble factors such as cAMP (49). tiation. We have extended earlier work (30) with the observation Transdifferentiation. Rastinejad and Blau (30) observed that the 3' UTR of skeletal a-tropomyosin evokes several complementation of an already committed cell type. Others aspects of the muscle program from fibroblasts. This strongly have observed changes from one cell type to another. This suggests that the tropomyosin 3' UTR is not merely comple- behavior, transdifferentiation, in which a cell type changes its menting some late step in muscle development but can turn on phenotype into that of another cell type or restricted set of cell the whole suite of genes involved in muscle development. To types, has recently been reviewed (44). Cell types involved our knowledge, transdifferentiation of nonmuscle cells into include pigmented retinal epithelial cells changing into neu- muscle by anything other than nuclear DNA binding proteins rons, pancreatic cells changing into hepatocytes, and muscle such as MyoD and myf has not been reported previously. cells changing into neurons and the multiple cell types present Most studies of transdifferentiation used embryonic cells in a tentacle (45-47). Cells capable of transdifferentiation are (44, 50); however, in some cases adult cells retain the ability to Downloaded by guest on September 27, 2021 7524 Cell Biology: L'Ecuyer et al. Proc. Natl. Acad. Sci. USA 92 (1995) 3. Greenberg, M. E., Shyu, A. B. & Belasco, J. G. (1990) Enzyme 44, 181-192. 4. You, Y., Chen, C. Y. & Shyu, A. B. (1992) Mol. Cell. BioL 12, 2931-2940. 5. Schiavi, S. C., Wellington, C. L., Shyu,A. B., Chen, C. Y., Greenberg, M. E. & Belasco, J. G. (1994) J. Biol. Chem. 269, 3441-3448. 6. Chen, C. Y., Chen, T. M. & Shyu, A. B. (1994) Mol. Cell. Biol. 14,416-426. 7. Bickel, M., Iwai, Y., Pluznik, D. H. & Cohen, R. B. (1992) Proc. Natl. Acad. Sci. USA 89, 10001-10005. 8. Bagby, G. C., Shaw, G., Heinrich, M. C., Hefeneider, S., Brown, M. A., DeLoughery, T. G., Segal, G. M. & Band, L. (1990) Prog. Clin. BioL Res. 352, 233-239. 9. Kennedy, I. M., Haddow, J. K & Clements, J. B. (1991) J. Virol. 65, 2093-2097. 10. Newman,T. C., Ohme-Takagi, M., Taylor, C. B. & Green, P. J. (1993) Plant Cell 5, 701-714. 11. Brown, B. D., Zipkin, I. D. & Harland, R. M. (1993) Genes Dev. 7, 1620-1631. 12. Pryhuber, G. S., Church, S. L., Kroft, T., Panchal, A. & Whitsett, J. A. (1994) Am. J. Physiol. 267, L16-L24. 13. Gorospe, M. & Baglioni, C. (1994) J. Biol. Chem. 269, 11845-11851. 14. Fyrberg, C. C., Labeit, S., Bullard, B., Leonard, K & Fyrberg, E. (1992) Proc. R Soc. London B 249, 33-40. 15. Spirin, A. S. (1994) Mol. Reprod. Dev. 38, 107-117. 16. Standart, N. & Jackson, R. J. (1994) Biochimie 76, 867-879. 17. Wormington, M. (1994) Bioessays 16, 533-535. 18. Gallie, D. R. & Walbot, V. (1990) Genes Dev. 4, 1149-1157. 19. Leathers, V., Tanguay, R., Kobayashi, M. & Gallie, D. R. (1993) Mol. Cell. FIG. 5. Controls. (A and B) Cells were transfected with a viral Biol. 13, 5331-5347. vector containing (A) or lacking (B) a tropomyosin 3' UTR insert and 20. Ueno, S., Kotani, Y., Kondoh, K, Sano, A., Kakimoto, Y. & Campagnoni, A. T. (1994) Biochem. Biophys. Res. Commun. 204, 1352-1357. stained for muscle tropomyosin; in the absence of insert (B), no cells 21. Zeyenko, V. V., Ryabova, L. A., Gallie, D. R. & Spirin, A. S. (1994) FEBS express muscle tropomyosin. (C and D) Cells transfected with plasmids Leu. 354, 271-273. encoding tropomyosin 3' UTR and ,B-galactosidase simultaneously 22. Sheets, M. D., Fox, C. A., Hunt, T., Vande Woude, G. & Wickens, M. and stained for tropomyosin (C) and 3-galactosidase (D); cells positive (1994) Genes Dev. 8, 926-938. for tropomyosin are also positive for ,B-galactosidase. (X680.) 23. Wharton, R. P. & Struhl, G. (1991) Cell 67, 955-967. 24. Seeger, M. A. & Kaufman, T. C. (1990) EMBO J. 9, 2977-2987. transdifferentiate (46, 51). In addition, transdifferentiation in 25. Berry, M. J., Banu, L., Harney, J. W. & Larsen, P. R. (1993) EMBO J. 12, 3315-3322. most cases involves the change of phenotype from one cell type 26. Dix, D. J., Lin, P. N., McKenzie, A. R., Walden, W. E. & Theil, E. C. (1993) to one other cell type, implying that a given cell type has some J. Mol. Biol. 231, 230-240. plasticity that is constrained as to the alternate cell types it can 27. Emery-Goodman, A., Hirling, H., Scarpellino, L., Henderson, B. & Kuhn, express. This is consistent with our observation that osteo- L. C. (1993) Nucleic Acids Res. 21, 1457-1461. to 3' UTR. 28. Kislauskis, E. H., Li, Z., Singer, R. H. & Taneja, K L. (1993) J. Cell Biol. blasts only respond partially the tropomyosin 123, 165-172. This system offers another possibility. Syngeneic mouse 29. Blowers, A. D., Klein, U., Ellmore, G. S. & Bogorad, L. (1993) Mol. Gen. myoblasts can be transplanted into and fuse with existing Genet. 238, 339-349. myofibrils; allogeneic cells are rejected in weeks, even with 30. Rastinejad, F. & Blau, H. M. (1993) Cell 72, 903-917. immunosuppression (52). The only syngeneic source of myo- 31. Rastinejad, F., Conboy, M. J., Rando, T. A. & Blau, H. M. (1993) Cell 75, 1107-1117. blasts for human transplantation is a patient's cells or those of 32. Isaacs, W. B., Cook, R. K, Van Atta, J. C., Redmond, C. M. & Fulton, an identical twin. Obtaining sufficient satellite cells to correct A. B. (1989) J. Biol. Chem. 264, 17953-17960. a gene defect and transplant them back into an individual will 33. Cho, Y.-J., Liu, J. & Hitchcock-DeGregori, S. E. (1990) J. Biol. Chem. 265, always be limited by the scarcity and inaccessibility of satellite 538-545. 34. Andersson, S., Davis, D. L., Dahlback, H., Jornvall, H. & Russell, D. W. cells. If postembryonic fibroblasts also transdifferentiate, this (1989) J. Biol. Chem. 264, 8222-8229. system will offer a new source of cells for transplantation. 35. Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Additionally, local treatment with the gene for skeletal tro- Smith, J. A. & Struhl, K (1992) Short Protocols in Molecular Biology pomyosin or its 3' UTR and a corrective gene might allow (Greene/Wiley, New York). 36. Hawley-Nelson, P., Ciccarone, V., Gebeyehu, G., Jessee, J. & Felgner, P. L. transdifferentiation of fibrotic cells in situ. (1993) Focus 15, 73-79. The mechanism by which the 3' UTR acts to transdifferen- 37. Lin, J. J.-C., Chou, C. S. & Lin, J. L.-C. (1985) Hybridoma 4, 223-242. tiate cells or to complement differentiation is unknown. It 38. Cripe, L., Morris, E. & Fulton, A. B. (1993) Proc. Natl. Acad. Sci. USA 90, seems unlikely, however, that the 3' UTR will bind directly to 2724-2728. may 3' more 39. Isaacs, W. B., Kim, I. S., Struve, A. & Fulton, A. B. (1989) J. Cell Biol. 109, DNA. It be that control through the UTR is 2189-2195. robust than regulation through DNA binding proteins such as 40. Fulton, A. B. & L'Ecuyer, T. (1993) J. Cell Sci. 105, 867-871. myoD, because small changes in the binding site of myoD can 41. Hansen, L. K, Mooney, D. J., Vacanti, J. P. & Ingber, D. E. (1994) Mol. alter the specificity of myoD to permit activation of other Biol. Cell 5, 967-975. genes. It is that the mechanism by which the 3' UTR 42. Rastinejad, F., Conboy, M. J., Rando, T. A. & Blau, H. M. (1994) Cell 75, possible 1107-1117. regulates gene expression will resemble the interactions be- 43. Choi, J., Costa, M. L., Mermelstein, C. S., Chagas, C., Holtzer, S. & tween tRNAs and charging enzymes, in which overall second- Holtzer, H. (1990) Proc. Natl. Acad. Sci. USA 87, 7988-7992. ary structure is more important than local sequences. If so, it 44. Eguchi, G. & Kodama, R. (1993) Curr. Opin. Cell Biol. 5, 1023-1028. seems likely that small mutations are likely either to be without 45. Kosaka, J., Watanabe, K & Eguchi, G. (1992) Exp. Eye Res. 55, 261-267. 46. Hall, P. A. & Lemoine, N. R. (1992) J. PathoL 166, 97-103. effect or to abolish activity. Thus, using the 3' UTR to induce 47. Schmid, V., Alder, H., Plickert, G. & Weber, C. (1988) Cell Differ. Dev. 25 muscle differentiation may entail less risk of accidently induc- (Suppl.), 137-146. ing expression of other, unwanted genes. Whether this proves 48. Bachem, M. G., Sell, K M., Melchior, R., Kropf, J., Eller, T. & Gressner, to be the case, the induction of a complete cell phenotype A. M. (1993) Virchows Arch. B 63, 123-130. 49. Lipton, B. H., Bensch, K G. & Karasek, M. A. (1991) Differentiation 46, solely through the expression of an RNA that produces no 117-133. protein opens exciting questions about gene expression. 50. Beresford, W. A. (1990) Cell Differ. Dev. 29, 81-93. 51. Bachem, M. G., Meyer, D., Schafer, W., Riess, U., Melchior, R., Sell, K M. 1. Laird-Offringa, I. A. (1992) Bioessays 14, 119-124. & Gressner, A. M. (1993) J. HepatoL 18, 40-52. 2. Herrick, D. J. & Ross, J. (1994) Mol. Cell. Biol. 14, 2119-2128. 52. Rando, T. A. & Blau, H. M. (1994) J. Cell BioL 125, 127-1287. Downloaded by guest on September 27, 2021